In general, in plants such as power plants, the state of a plant is monitored to prevent a state such as an accident or a stop in operation. To monitor a plant, for example, points to be monitoring objects included in monitoring items in a plant (e.g., called a PID or TAG) are defined. Furthermore, the state of a plant is monitored by comparing an upper limit value and a lower limit value as thresholds that are set for a point in a monitoring object with a process value acquired from the point in the monitoring object. In this case, the judgment of the presence of an anomaly is made by judging whether the process value exceeds the upper limit value or the lower limit value.
In addition, in order to improve judgment accuracy, there have been introduced techniques to improve the search accuracy of judgment by calculating these upper limit value and lower limit value using the average value, the standard deviation, and the like of past plant data in a monitoring point. Furthermore, a method has been used in which a main variable point for a point to be monitored (a point having a strong correlation) is found, and under a correlation with the main variable point, an upper limit value and a lower limit value from which the operating state or the variations in a plant are eliminated are set to perform monitoring. For example, one such method is to find a main variable point for a monitoring point, divide a monitoring zone into some clusters, and calculate a monitoring threshold using a weight coefficient with respect to the number of data items of the monitoring point to perform monitoring.
However, although some points can be monitored in such a method, it has been known that some monitoring points cannot be accurately monitored by these method of cluster division, method of calculating monitoring threshold, and method of using weight coefficient.
In addition, it has been known that some monitoring points have no main variable points. That is, there are monitoring items that cannot be monitored using only a correlation with the main variable point, and a new monitoring method that supports the characteristics of monitoring items is needed.
In addition, the investigation of appropriateness or validity of thresholds and a new monitoring method that supports the characteristics of monitoring items are performed while the operation of a plant is continued. For this reason, it takes a long time to investigate the appropriateness or validity of effective thresholds or the new monitoring method.
In addition, a plant monitoring system is required not only to monitor the state of a plant that reaches an alerting zone but also to monitor the state of a sign or a precursor before reaching the alerting zone, as a monitoring item.
The present invention thus has an object to provide a plant monitoring system by which the accuracy in state monitoring of a plant can be enhanced.